Welcome to the
Carnegie Mellon University's home-page
for the Q-MARS project, which is part of a MURI effort with the University
of Illinois at Urbana-Champaign and the University of Virginia at
Charlottesville. This MURI effort focuses on QoS Support for
Surveillance and Control Systems, and is funded by DARPA and
ONR.
In the surveillance and control systems, Quality-of-Service support is increasingly important to provide real-time
results under an dynamic environment with uncertainty. In many applications the relation between level
of service and resource requirements is not fixed. Environmental factors
outside the control of the system affect this relationship and may also
affect the perceived utility for a given level of service. An example of
such application is radar tracking.
In order to guarantee optimal timing services with limited
resources (including contented resources, limited priority
granularity, etc.), the following research problems need to be
solved: resource management with environmentally dependent measures
of quality, optimization with indirect mapping between QoS and
application control, prediction of performance with stochastic
tasks, management of heterogeneous resources including both
computing and physical resources, development of protocols for QoS
control of applications and mechanisms. We divide the above
problems into Q-RAM
(QoS-based Resource Allocation in Real-time Distributed System),
Quantized Earliest-deadline-first scheduling.
We are implementing these solutions under the framework of Sesco
(Session-Coordinator) system.
Q-MARS